Sensitivity to surrounding circumstances is essential for the safety of mountain scrambling. In this paper, the authors present a smart helmet prototype equipped with visual SLAM (simultaneous localization and mapping) and barometer multi-sensor fusion (MSF), IMU (inertial measurement unit), omnidirectional camera, and global navigation satellite system (GNSS). They equipped the helmet framework with SLAM to produce 3D semi-dense pointcloud environment maps, which are then discretized into grids. Then, the novel danger metrics they proposed were calculated for each grid based on surface normal analysis. The A* algorithm was applied to generate safe and reliable paths based on minimizing the danger score. This proposed helmet system demonstrated robust performance in mapping mountain environments and planning safe, efficient traversal paths for climbers navigating treacherous mountain landscapes.